#BigData allows #insurance companies new ways to reduce their operational risk by providing a 360 degree view of the customer.
Click to Tweet.

Big Data: Hidden Opportunities

The 20th Asian Actuarial Conference was held in Gurgaon India. The event’s theme was Changing Asian Societies: Challenges and Opportunities. Kaushik Mitra, Senior Vice President, Actuarial, Data Science & BIR, AXA Business Services, spoke at a panel discussion on Big Data at this event. Here are the excerpts from his discussion.

What is the current state of big data and applications in insurance?

Big Data today allows Insurance companies new ways to reduce their operational risk by having a 360 degree view of the customer. Integrated with social media analytics, Big Data helps to quickly ascertain risk profile of current customers and prospective clients and hence price risk more accurately. Overall there is convergence of data points that have an interesting story to tell. Also the rise of digital wearable devices is completely revolutionizing this space with the availability of real-time data.

Telematics in motor insurance, wearable tech in life and health insurance, intelligent IOT enabled devices for property- technology is aiding the development of disruptive products/services and driving operational efficiencies. The analysis of customer activity and feedback across social media, telephone calls, email, chats, wearable devices and intelligent gadgets is influencing important decisions in the insurance landscape on the sales, cross sell, up-sell, customer retention and customer experience dimensions. Data collected can be segmented on various parameters and those with high propensity to convert can be targeted. For instance data will tell you if a segment of customers in the non-life insurance space are good prospects for P&C insurance too. The industry is also leveraging big data to detect extremely sparse events.

At AXA we are also broadening the scope of big data and going beyond identifying new customers, cross sell and up-sell to predicting fraud, proactively identify claims that could potentially have large impact on cost and leveraging the analysis for capital reserve optimization. We are very well prepared to extract value from data with our dedicated in house Data Innovation Lab.

What are the differences between traditional analytics and Big Data and machine learning approaches?

There are a couple of fundamental differences.

First, in the world of Big data there is large abundance of data which may not be clean, perfect to start with. While sifting through this and analyzing can yield new insights that have not been explored before and will help us insure things which were hitherto difficult to insure, this will also require change in mindset to move from a completely explainable equation to a more convoluted black-box model. For instance it can give us information on people who don’t have a driver’s license. This becomes very important as we decide on how to underwrite the risk.

Second is the difference between hypothesis driven approach of the past and data driven approach in the current environment. The hypothesis approach considerably narrows the learning process and the outcome from the beginning. When guided by the hypothesis it makes it very difficult to discover any other possible explanations. In the data driven approach, we let data guide us. This throws open several new insights, for example it can help better in overall risk estimation or fraud detection.

What are some of the hidden opportunities in Big data in the insurance industry in the next five years

The next big change that will completely transform the insurance landscape is robo-advisors. Robot advisory will help provide more intelligent, real time advice going beyond price to services like validation like suggesting a reassessment of needs based on market situation and proactively preparing clients in anticipation of extreme risks within a region/area—like changing weather conditions, probability of flood and so on. While the advantages to consumers are numerous, AI advisory can also keep the insurance industry protected against fraudulent events by tracking claims activity, analyzing patterns and recommending actions.

Another key change we will experience in the industry is the adoption of deep learning methodologies. For instance deep learning will help insurance companies assess layers of data in a motor accident scenario and identify a fraud pattern. It can also help insurance companies provide proactive customer experience. For instance in health insurance, deep learning can help predict when the next checkup is required.

Big Data: Hidden Opportunities

The 20th Asian Actuarial Conference was held in Gurgaon India. The event’s theme was Changing Asian Societies: Challenges and Opportunities. Kaushik Mitra, Senior Vice President, Actuarial, Data Science & BIR, AXA Business Services, spoke at a panel discussion on Big Data at this event. Here are the excerpts from his discussion.

What is the current state of big data and applications in insurance?

Big Data today allows Insurance companies new ways to reduce their operational risk by having a 360 degree view of the customer. Integrated with social media analytics, Big Data helps to quickly ascertain risk profile of current customers and prospective clients and hence price risk more accurately. Overall there is convergence of data points that have an interesting story to tell. Also the rise of digital wearable devices is completely revolutionizing this space with the availability of real-time data.

Telematics in motor insurance, wearable tech in life and health insurance, intelligent IOT enabled devices for property- technology is aiding the development of disruptive products/services and driving operational efficiencies. The analysis of customer activity and feedback across social media, telephone calls, email, chats, wearable devices and intelligent gadgets is influencing important decisions in the insurance landscape on the sales, cross sell, up-sell, customer retention and customer experience dimensions. Data collected can be segmented on various parameters and those with high propensity to convert can be targeted. For instance data will tell you if a segment of customers in the non-life insurance space are good prospects for P&C insurance too. The industry is also leveraging big data to detect extremely sparse events.

At AXA we are also broadening the scope of big data and going beyond identifying new customers, cross sell and up-sell to predicting fraud, proactively identify claims that could potentially have large impact on cost and leveraging the analysis for capital reserve optimization. We are very well prepared to extract value from data with our dedicated in house Data Innovation Lab.

What are the differences between traditional analytics and Big Data and machine learning approaches?

There are a couple of fundamental differences.

First, in the world of Big data there is large abundance of data which may not be clean, perfect to start with. While sifting through this and analyzing can yield new insights that have not been explored before and will help us insure things which were hitherto difficult to insure, this will also require change in mindset to move from a completely explainable equation to a more convoluted black-box model. For instance it can give us information on people who don’t have a driver’s license. This becomes very important as we decide on how to underwrite the risk.

Second is the difference between hypothesis driven approach of the past and data driven approach in the current environment. The hypothesis approach considerably narrows the learning process and the outcome from the beginning. When guided by the hypothesis it makes it very difficult to discover any other possible explanations. In the data driven approach, we let data guide us. This throws open several new insights, for example it can help better in overall risk estimation or fraud detection.

What are some of the hidden opportunities in Big data in the insurance industry in the next five years

The next big change that will completely transform the insurance landscape is robo-advisors. Robot advisory will help provide more intelligent, real time advice going beyond price to services like validation like suggesting a reassessment of needs based on market situation and proactively preparing clients in anticipation of extreme risks within a region/area—like changing weather conditions, probability of flood and so on. While the advantages to consumers are numerous, AI advisory can also keep the insurance industry protected against fraudulent events by tracking claims activity, analyzing patterns and recommending actions.

Another key change we will experience in the industry is the adoption of deep learning methodologies. For instance deep learning will help insurance companies assess layers of data in a motor accident scenario and identify a fraud pattern. It can also help insurance companies provide proactive customer experience. For instance in health insurance, deep learning can help predict when the next checkup is required.

Leading Machine Learning and Artificial Intelligence Cultural Change

The first AXA Chair in Greece explores epigenetics

The impact of demonetization on the insurance industry

Smart Cities: Step into the city of the future!